Introduction

Prostate cancer is one of the most common malignancies in men worldwide. Prostate-specific antigen (PSA) is presently the most widely used tumor marker for prostate cancer diagnosis, staging and treatment follow-up [1, 2]. PSA levels are expected to increase with age as a result of hypertrophied prostate. Therefore, using the current standard single cut-off of 4.0 ng/ml, might not be appropriate for all age groups. Age-adjusted PSA levels might help in increasing the sensitivity rate of prostate cancer detection in younger men and in decreasing unnecessary biopsies in older subjects [3–6].

In 1993, Oesterling et al. [7] first introduced age-specific reference ranges for serum PSA for Caucasian Americans. However, the rate of prostate growth and the volume of prostate-producing PSA is known to vary among ethnic groups [8, 9], and Oesterling’s age-specific PSA reference ranges may not be generally applicable to all populations.

Many studies about age-adjusted PSA values have been conducted for African Americans [10], Koreans [11], Japanese [8], and Chinese [12–14], and even for Taiwanese [15], but no large-scale study has been conducted among Taiwanese.

This study determines age-specific PSA distributions in Taiwanese men, with the goal of offering optimal values for detecting the upper limit of serum PSA levels based on a large study cohort.

Methods

This study recruited 213,986 Taiwanese men aged above 19 years old without history of prostate cancer, urinary tract infection, or prostate infection who had undergone a routine health checkup from the MJ Health Screening Center in Taiwan from January 1999 through December 2016 [16]. The MJ Health Database only includes participants who provided informed consent. The study protocol was approved by the Ethics Committee of Taichung Tzu Chi Hospital, Buddhist Tzu Chi Medical Research Foundation (REC 109–54).

Participants were divided into seven age cohorts based on age at the time of the most recent screening: 20–29 years of age (6,382 individuals), 30–39 (33,555), 40–49 (78,912), 50–59 (50,229), 60–69 (31,252), 70–79 (13,152), and above 79 (504).

The PSA values were analyzed using Stata version 14.2 (Stata, College Station, TX, USA). The association between serum PSA levels and age was measured using Pearson product-correlation coefficients [17]. Descriptive statistics including the mean, median, 5th, 25th, 75th, and 95th percentiles of the PSA level distribution were calculated for each 10 year age group cohort [18].

Results

Among the 213,986 participants, the mean PSA level was around 1.21ng/ml. Table 1 shows the mean serum PSA concentration and median, 5th, 25th,75th and 95th percentiles for each age cohort. The median PSA value (95th percentile range) was 0.70 ng/ml (1.80) for men 20–29 years old; 0.70 ng/ml (1.80) for men 30–39 years old; 0.70 ng/ml (2.0) for men 40–49 years old; 0.80 ng/ml (3.2) for men 50–59 years old; 1.10 ng/ml (5.60) for men 60–69 years old; 1.50 ng/ml (7.40) for men 70–79 years old and 1.60 ng/ml (8.40) for men over 79 years old Table 1. The serum PSA values were correlated with age (r = 0.274, p<0.001; 95% CI 0.270 to 0.278).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Serum PSA according to age in 213,986 Taiwanese men.

https://doi.org/10.1371/journal.pone.0283040.t001

The results show only small differences in the median and 95th percentile PSA values among men younger than 50 years of age, but significant increases were found at higher ages. In those aged 20–59, 1.48% (2,510/169,078) had serum PSA levels greater than 4.0 ng/ml, as opposed to 11.38% (5,112/44,908) of men above 59 years old. Table 2 summarizes the proportions of men with various PSA values based on age.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Proportion of men with various serum PSA levels according to age.

https://doi.org/10.1371/journal.pone.0283040.t002

Discussion

While PSA testing has been used for decades for detecting early prostate cancer and monitoring therapy response, PSA levels are significantly affected by age, hypertrophy of the prostate, environmental factors, geography, diet and ethnicity. Thus, the age-specific PSA reference ranges proposed by Oesterling et al. [7] might not be applicable across all countries and ethnic groups. Studies from some Asian countries such as Korea, Japan and China suggest that, for all age groups, serum PSA levels in Asian people are lower than in other ethnic groups. However, Lin et al. [15] found that serum age-specific PSA levels among Taiwanese men are higher than those in other Asian ethnicities in all age groups, and approximate the levels in African-American men and Caucasian American men. The current study seeks to determine optimal age-specific PSA reference ranges for Taiwanese men based on a large cohort.

In this study, the median and 95% percentiles of serum PSA level in Taiwanese men are found to be directly related to age, and rise rapidly over the age of 50, consistent with results for different ethnic groups in different countries. Hypertrophy of the prostate can account for the elevation of serum PSA levels with age. However, over the age of 60, our subjects have higher age-specific PSA reference ranges than those observed by Oesterling et al. [7], which is contrary to the results for Japanese, Korean and Chinese subject Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Comparison of serum PSA levels among African-Americans [21], Caucasian Americans [7], Asian-Americans [21], Taiwanese [15], Japanese [8], Koreans [11], and Chinese [12–14].

https://doi.org/10.1371/journal.pone.0283040.t003

When compared with the age-specific PSA reference ranges from Oesterling et al., our results show that the difference in age-specific PSA reference ranges are relatively small for the 40-49-year-old age group (2.0 vs 2.5) and the 50-59-year-old age group (3.2 vs 3.5), but the direction of difference is inverted and obvious in the 60–69 (5.6 vs 4.5) and 70–79 (7.4 vs 6.5) year-old age groups Table 3. As a consequence, using the age-specific PSA reference ranges proposed by Oesterling et al. for screening of prostate cancer in Taiwan might result in errors by increasing sensitivity and decreasing specificity in patients over 60 years old. On the other hand, for those under 60 years old, errors in prostate cancer screening would still result from increased specificity and decreased sensitivity. From this study, the age-specific PSA reference ranges can be deduced to enhance the detection rate for prostate cancer in younger patients with PSA levels below the single cut-off of 4.0ng/ml and thus reduce the number of negative biopsies in older men with PSA levels > 4.0ng/ml [19]. Reissigl et al. [20] reported that using the age-specific PSA cut-off proposed by Oesterling et al. resulted in an 8% increase in the number of biopsies and an 8% increase in organ-confined cancer detection in men aged 45–59 years, while in older men aged 60–75 years, it resulted in about a 21% reduction in biopsies but would result in 4% of prostate cancers being missed.

Whether our results for age-specific PSA reference ranges are optimally appropriate for Taiwanese men is still open to question and the rate of accurate detection for prostate cancer remains to be determined. Further rigorous studies will be needed with tools including digital rectal examinations, transrectal ultrasound and prostate biopsy.

Our present results support the findings of a previous study for Taiwanese subjects by Lin et al [15], in which increasing age is found to produce higher PSA reference ranges than those of other Asian countries and near to the levels found in African-American men. Intriguingly, despite common ethnical origins with Chinese subjects, we note that serum age-specific PSA levels among men in Taiwan are higher than those in various regions of mainland China. The reason for this discrepancy remains unclear, though dietary and environmental factors may play a role [13].

The present study is subject to certain limitations. First, prostate biopsy, or transrectal ultrasound was not performed in participants with PSA levels > 4.0 ng/ml or abnormal digital examination. Thus the cohort could include prostate cancer cases, leading to increased PSA level reference ranges. Nevertheless, the large size of the cohort used in this study could reduce statistical bias. Second, most clients of the MJ Health Screening Center have a relatively high levels of disposable income, and thus lower economic class males may be underrepresented in the cohort.

Conclusion

Serum PSA concentration increases with age, and the reference ranges of age-specific PSA levels in Taiwanese men are different from those in other ethnic groups and also in among ethnic Chinese in mainland China. The age-specific PSA reference ranges for Taiwanese men aged 60 and over are higher than those proposed by Oesterling et al. that are currently used in Taiwan and higher than those of other Asian countries. Age-specific PSA reference ranges should be population-specific. Our results may offer the important data to determine the age-specific PSA reference ranges in different age groups among Taiwanese men.

Acknowledgments

All of the data used in this research were authorized by, and received from MJ Health Research Foundation (Authorization Code: MJHRF2020001A). Any interpretation or conclusion described in this paper does not represent the views of MJ Health Research Foundation.